Ahmed Raza Ur Rehman

Developed an LLM-based AI agent framework with RAG pipelines, vector memory, tool usage, and multi-agent orchestration. Designed workflows for AI reasoning, retrieval, and automated task execution. Evaluated AI-generated outputs and improved system reliability through structured testing.

• Technologies: Python, LangChain, RAG, FastAPI, ChromaDB, OpenAI APIs

Designed a custom 5-stage pipelined RISC-V processor with an integrated neural network acceleration unit. Implemented pipeline control, hazard handling, and matrix multiplication acceleration. Created technical documentation covering architecture decisions, verification strategy, and system behavior.

• Technologies: SystemVerilog, RISC-V, FPGA, Vivado, QuestaSim

Developed an AI-powered intrusion detection system for identifying network attacks and anomalies. Applied machine learning techniques including Random Forest, XGBoost, and LSTM Autoencoders. Analyzed network traffic patterns and evaluated detection performance.

• Technologies: Python, Scikit-learn, XGBoost, TensorFlow, Wireshark, Docker

Built an optimized real-time object detection system using YOLOv8, TensorFlow Lite, and FPGA acceleration. Applied model optimization techniques including INT8 quantization. Designed solutions for AI inference on resource-constrained devices.

• Technologies: Python, TensorFlow Lite, OpenCV, FPGA, ESP32